Frozen section discrepancy in the evaluation of central nervous system tumors.

CONTEXT Frozen section (FS) evaluation of central nervous system (CNS) lesions provides an assessment of specimen adequacy and facilitates triage for ancillary studies. Frozen section also provides an accurate preliminary diagnosis; however, certain lesions are recognized to cause diagnostic challenges at FS. OBJECTIVE To identify cases in which there was a discrepancy between the FS diagnosis and final diagnosis in the case to heighten awareness of common diagnostic pitfalls in surgical neuropathology. DESIGN All CNS FS cases involving a tumor diagnosis at FS or permanent section (N = 2156) from September 1997 until June 2005 were retrospectively reviewed. Discrepancies between the FS and final diagnoses were identified. RESULTS Of the 2156 cases identified, 57 (2.7%) discrepant diagnoses were found. Twelve (21.1%) of 57 discrepancies involved errors in classification of spindle cell lesions, most commonly confusing schwannomas or meningiomas with other lesions. Twelve (21.1%) of 57 cases involved errors in differentiating oligodendrogliomas from astrocytomas. Nine (15.8%) of 57 discrepancies involved errors in the diagnosis of CNS lymphoma. Eight (14.0%) of 57 cases involved errors in differentiating reactive from neoplastic processes, most frequently gliosis versus glioma. Four (7.0%) of 57 discrepancies involved errors in the overgrading of tumors. The remaining 12 (21.1%) of 57 cases included an assortment of other discrepancies. CONCLUSIONS Frozen section of CNS neoplastic processes can be highly accurate. Less than 3% of FS diagnoses in 1 institution's experience were discrepant with the final diagnoses. Approximately 80% of the discrepant cases were classified into 5 categories: spindle cell lesions, astrocytoma versus oligodendroglioma, differential diagnosis of CNS lymphoma, reactive versus neoplastic process, and tumor overgrading. Awareness of these pitfalls may help in further increasing diagnostic accuracy.